OpenAI must face a lawsuit from news publication The Intercept alleging the tech giant removed copyright management information from articles to train its AI models. A judge explained on Thursday that these actions “implicate(d) the same kind of property-based harms traditionally actionable in copyright.”
Allegations of violating the DMCA and using content for ChatGPT training
In February 2024, The Intercept sued OpenAI for removing the copyright management information from its articles, violating the Digital Millennium Copyright Act, and then using them as ChatGPT training data without permission. The news outlet also alleged that Microsoft, as a significant investor, was complicit.
A few months later, both tech companies argued the lawsuit should be dismissed because The Intercept did not adequately demonstrate harm and AI-generated outputs did not count as reproductions. In November, Judge Jed Rakoff made the decision to dismiss Microsoft from the case and the claim that OpenAI unlawfully distributed articles.
However, the DMCA claim about the intentional removal of copyright information was allowed to proceed, and on Feb. 20, 2025, the judge explained his reasoning behind that decision. According to court documents uploaded by ChatGPT is eating the world, OpenAI should have known that removing the copyright information would be regarded as infringement.
He dismissed the claim that OpenAI’s removal of copyright information from training articles led to ChatGPT users producing “regurgitations” of copyrighted works without attribution due to insufficient evidence actually linking the company to their distribution. The Intercept also did not adequately prove Microsoft was involved in any of its accusations.
“The DMCA adds another stick to the bundle of property rights already guaranteed to an author in her work under traditional copyright law,” Judge Rakoff wrote. He added that, despite the modern application of the DMCA, the violation produced “the same kind of harm long recognized in copyright suits.”
The decision goes directly against the ruling of a similar case from Nov. 7, where Raw Story and AlterNet, two other independent news outlets, argued that OpenAI violated the DMCA by removing copyright management information from their articles. In that case, Judge Colleen McMahon dismissed the lawsuit, finding that the plaintiffs failed to demonstrate sufficient harm.
Matt Topic, The Intercept’s attorney, told Reuters his law firm Loevy + Loevy “look(s) forward to the ultimate vindication of our rights.” An OpenAI spokesperson also reinstated that its models are “trained on publicly available data, grounded in fair use and related principles that we view as fair for creators, and support innovation.”
Growing concerns about journalism and AI companies
How publications can coexist with AI companies is still a point of contention in the journalism industry. Many outlets have opted to sell their content to the tech giants to train their models, such as The Guardian, Associated Press, Reuters, LA Times, The Independent, AFP, News Corp, Financial Times, Time, Conde Nast, and The Atlantic.
There are growing concerns that the strategy will be short-lived, and news outlets will ultimately struggle to survive if their content appears in AI summaries instead of generating direct reader traffic to their websites.
Some outlets are choosing to beat them rather than join them, suing the tech giants for unauthorised use of their content for AI training. The New York Times is one of them, and has been embroiled in a legal battle with OpenAI and Microsoft since 2023. This month, Conde Nast and a number of other major publishers have accused AI startup Cohere of similar copyright infringement.